Low mass galaxies seen by MUSE and gravitational lensing

Galaxy formation and evolution is one of the most challenging mysteries in the observable Universe. In order to improve our knowledge in this field, the research make use of different observation programs to characterize scaling relations of physical properties, to better constrain and understand galaxy population at different stages of their lives throughout the history of the Universe. Since the past decades, studies are trying to extend those scaling relations in the parameter space. It is in this movement that this thesis project fits. The Spectral Energy Distribution (SED) of high redshift galaxies contains the signature of physical properties such as stellar mass, SFR and extinction. In this work, we perform a SED analysis of magnified galaxies at z>3 using deep Hubble, VLT and Spitzer/IRAC images of the Frontier Fields galaxy clusters. Due to the size of the Kband PSF and specially IRAC PSF and the high density of bright cluster members, it is crucial to deblend Kband and IRAC images to get a reliable SED at longer wavelengths. We do this by automatically fitting the contaminating galaxies with GALFIT, using a custom Python script which accounts for the relative levels of contamination from each cluster member. We model the decontaminated SED using stellar population models. We apply this method to derive SFR, masses and sizes of a sample of 63 galaxies at z > 3 detected in the Abell 2744 and MACS0416 fields, spectroscopically confirmed with MUSE. The very strong amplification of these clusters allow us to collect a robust sample of low-mass / low-luminosity galaxies (∼ 10⁸ M), probing the low-luminosity part of scaling relations between stellar mass and size and stellar mass and stellar formation rate. (author)